This was my final project for the "How to make (almost) anything" class at the [MIT Media Lab][21], when I was a first-year MSc. student. As my field of research is technologies for learning, I am interested in ways to make the opaque transparent, to make people understand what is "under-the-hood". Computers are everywhere, but we hardly know how they work. So this was the motivation for this project (if you are interested in this topic, I also wrote and directed a documentary on how microchips and computers work - have a look [here][18] and email me if you want a copy). If you are interested in education and new technologies, please check out these [cool videos][22], other projects on my [website][23], the [Gogo Board][24] website, or the [NetLogo][25] website.
The idea of the project was to build a device that could do computation without electrons (well, not considering the electrons in water itself). Water was a interesting choice, in fact, _Fluidics _is a very important field of study that is widely used in aerospace or mission-critical applications, where electronic control devices don't offer the needed reliability or cannot support the environment. Also, Fluidics has been use in military equipment in order to prevent malfunction in a nuclear war, when electric devices cease to work.
However, the idea was not to send people to space or to control missiles, but rather make a device that could help people build computation with their own hands, demystifing the computer. To do computation, a great thing to have are obviously boolean operations: but how to implement them with water?
After many hours of research on binary logic, a lot of sketching and unsuccessful attempts, I decided to try this to use the laser cutter and build something with multiple layers. A simple logic gate looks simple on paper, but water is a messy media to work with... The final assembled adder looked like this:
![][26]
This a zoom-in of one of the modules:
![][27]
It works based on one very simple principle: there are two inputs of water on the top.
| ----- |
| When just one of them is "on"... | When both are "on", the two jets collide, going vertically down to the "U" piece, that collects the water. |
| ![][28] | ![][29] | |
| In other words, it's a XOR gate.
| In other words, an AND gate.
|
So, we have two Boolean operations in just one device: AND and XOR. So, that's a half-adder!
![][30]
But it was a long way to get there. My first trial looked like this:
![][31]
There was leakage all over, so I tried a new design:
![][32]
It worked better, but its size could not cope with the flow of water. However, the principle worked very well.
![][33]![][34]![][35]
This was the third trial, and it worked fine.
Then I fabricated many of them, using the laser cutter and acrylic of various colors. I glued a Lego piece under the device, so that it would be easier to assemble small "computers". This is how the 4-bit adder looked like, after many (I mean, many!) hours of wet work. If you think programming a computer is hard, just imagine what it would be if your bits were leaking all over the place.
![][36]
![][37]
![][38]![][34]![][39]
[Just a quick note: I'm well aware that there has been tons of research on Fluidics. But this was just a class project, and it was never published, so that's why you will see no references. This is an [old version][40] of this page]